By Mike Stringer
Citrix recently formed the Data Innovation Group (DIG), which consists of three teams that are all part of Citrix’s IT Organization; Information Management, BI and Data Science. The team’s underlying core function is to make data and insights available, enabling data driven intelligence for both our company and our customers.
Citrix products sit on top of a very rich strip of customer and product data - few other companies are positioned at the intersections of backend servers, applications, networks, mobile devices, meetings and people. We have a tremendous opportunity to collect and analyze this naturally occurring data. This data is instrumental in helping our customers recognize more value from our products and for Citrix to better understand our customers and service them better.
One of the top priorities of the DIG team’s journey to harness data to drive customer intimacy was to develop and host a customer facing, centralized data upload service. We developed the service in Amazon’s AWS environment to leverage the scaling, storage and analysis capabilities (S3, PostgreSQL, EMR, and RedShift) the AWS Cloud offers. We leveraged our Citrix NetScaler VPX AWS offering for load balancing and SSL offloading as well as our Cloud Bridge VPX offering to seamlessly connect the AWS and Citrix networks. Today, customers are able to send (both manual and phone home) Citrix product and usage data to this upload service. As a value exchange, Citrix analyzes that data and provides information back to customers in the form of known problems, best practices and other recommendations (Citrix Insights Services, previously Tools as a Service (TaaS)). A majority of Citrix’s enterprise products leverage this service with over 600 GB of customer data being uploaded per week.
The customer data is also analyzed with Amazon EMR and Hadoop and mashed up with a number of other internal data sources to help Citrix improve our relationship with Customers. One of the interesting customer data areas the DIG team is working on is the ability to better predict at-risk customers. For example, at-risk customer support cases, at-risk customer accounts, at-risk account renewals and at-risk sales opportunities.
Using machine learning algorithms like Random Forest, the team is able to identify and then proactively engage with these at-risk customers as early as possible, helping improve the overall customer experience, driving higher loyalty ratings, and ultimately, improving customer lifetime value. However, not all at-risk scenarios can be handled or managed in the same way.
To enable a more action-oriented approach, we augment the predictive results with prescriptive information based on the attributes that drove each prediction. For example, many attributes go into an at-risk prediction,
however all those attributes are not actionable. We can’t control how long a customer has been a customer, and we can’t take action on how many employees a customer has, even though both are great predictive attributes. However, we can take action on other attributes like, number of sales interactions, number of critical support cases, time cases are open and number of product problems. The prescriptive approach allows teams to focus resources on the at-risk scenarios that are most likely able to be helped and helps teams decide what actions are likely to drive desired outcomes.
Dark Data also plays an increasingly important role across the Citrix organization. Citrix defines dark data as data that is either very difficult or impossible for people to analyze manually and therefore of little value without advanced analytics.
We have developed a text analytics platform that uses a number of open source libraries and Natural Language Processing (NLP) capabilities (including Stanford’s NLP) to help create deeper insights and attributes from free form text. These new sources for attributes can also be used in predictive models like the at-risk models as well as for understanding and taking action on insights from support cases, forums, and free text from surveys like NPS and Sales Win/Loss surveys. We even use these text analytics capabilities to better understand the free text from our internal global employee survey, helping improve our employees’ experience as well. We believe this approach to text analysis is interesting because the data is allowed to “speak for itself “and that uncovers patterns, trends and insights that were previously unknown and not being asked. Madhav Chinta, our data science director often states, “…we don’t know what we don’t know.” The resulting text analytics tools and capabilities also have far reaching value across the company including helping the Product Management and Product Engineering teams make product decisions as well as delivering value for the Sales and Services teams.
It’s important to note that the Data Innovation Group has built a team with a core competency in data and advanced analytics, end to end. However, partnering with the business teams within Citrix is critical as DIG is not responsible for and doesn’t always have the business experience to make customer and product decisions or for improving areas like product quality and enhancing customer experience.
The focus and responsibility of the group is to make data more readily available and increase the ease with which employees , partners and customers (via data products) can access and consume data and insights. Access to these insights enables key groups across Sales, Services and Product teams to make better decisions based on data.
We also strongly believe that data driven decisions are an “AND” opportunity versus an “instead of” opportunity. You’ll often hear people stating that data and advanced analytics can help take the intuition and emotion out of decision-making. However, Citrix believes that by leveraging advanced analytics to create insights and simplifying complex data sets, decisions makers are able to use the data and insights to augment their experience, intuition and emotions to enable better decision making.
The overall objective of our data driven intelligence initiatives is to support management decisions that drive better relationships with Citrix customers and improve the overall quality of Customer experiences when interacting with our products and services. You can find more information about Customer Insights Services here.